Adverse drug reactions (ADRs) are one of the leading causes of hospitalization and death in the United States. ADRs are often associated with unfavorable drug bioavailability or biodistribution profiles. Thus, ADRs could be prevented by optimizing drug transport properties -from the systemic, organ level down to the microscopic, cellular level. To improve the quality of drugs entering clinical trials, a new generation of microscopic imaging instruments -known as "high content screening" or "HCS" systems has been developed. HCS instruments can provide preclinical, human cell-based data to complement animal studies in predictive toxicology testing. As a high-throughput platform, HCS systems can be used to screen large collections of small molecules in physiologically-relevant assays. Now the challenge is to incorporate HCS technology into standard biomedical research practice, to facilitate discovery of less toxic drug candidates with improved clinical success rates. To meet this challenge, we propose to develop a cheminformatic and image data management and analysis plan to study the subcellular localization of fluorescent, small molecules -in living cells. Inspired by machine vision approaches currently being used as a tool to analyze the subcellular distribution of proteins on a genome-wide scale ("location proteomics"), we propose that machine vision could also be adopted as a tool to analyze the distribution of small molecule fluorescent drug candidates. In analogy to how protein location is encoded by signal peptides, we hypothesize that subcellular small molecule localization is encoded by "Chemical Address Tags" to be discovered within the chemical structure of small molecules. To test this hypothesis, we plan to: 1) Develop automated, image analysis and cheminformatic tools to reverse- engineer Chemical Address Tags in an objective, quantitative and high-throughput manner; 2) Develop and compare two quantitative, machine vision approaches to assay the transport properties of mitochondria- targeting molecules; 3) Demonstrate how a cheminformatics-driven, image data management and analysis plan can impact an anticancer drug lead optimization effort. [unreadable] [unreadable] [unreadable] [unreadable]